Determine the historical decision-evaluator version for ChainBench E1/E1B

Determine the exact historical source version of the decision evaluator used to generate the archived ChainBench E1/E1B decisions.

Background

Archived ChainBench decisions and derived statistics can be reproduced, but the source digest of the decision evaluator was not recorded. Consequently, numerical agreement does not establish that the evaluator implementation used historically is known or exactly identified.

References

The archived run artifacts do not record the source digest of the decision evaluator. Matching decisions and data digests therefore establishes the cohort correspondence but not the exact historical evaluator code version.

— Beyond Success Rate: Interventional Evaluation of Memory in Vision-Language-Action Systems  (2610.23372 - Luo et al., 16 Oct 2026) in Appendix B, Section “ChainBench Construction and Retrieval Diagnostics” and Section “Memory Maintenance at the Decision Level”